Ind. Eng. Chem. Res. 1994,33, 1468-1475
1468
Photodecomposition of Chlorine Dioxide and Sodium Chlorite in Aqueous Solution by Irradiation with Ultraviolet Light H e r d Cosson and William R. Ernst* School of Chemical Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0100
Rates of reaction and product formation were measured in photodecomposition experiments of aqueous sodium chlorite at 253.7 nm in a reactor that was continuously sparged with nitrogen to remove chlorine dioxide. Rapidlyremoving chlorine dioxide greatly reduced the formation of chlorate. The results of this work suggest that chlorate is not formed by direct decomposition of chlorite, but rather by decomposition of chlorine dioxide. The results are consistent with the stoichiometry, 3C102- H20 (+hv) C1- 2C102 20H- 0.50,. The rate of photodecomposition of sodium chlorite was studied over a pH range of 4-10 and a t unbuffered conditions. Distribution of major products was not affected by pH; rates of reaction and chlorine dioxide formation were maximum a t p H 6. Results of this work provide indirect evidence of the validity of certain elementary steps in mechanisms that have been proposed by previous workers. Quantum yields were measured for both photodecomposition of chlorine dioxide and sodium chlorite. For the former reaction, the values were 0.44 a t 253.7 nm and 1.4 a t 300 nm. For the latter reaction, the values a t 253.7 nm ranged from 0.72 to 1.53, depending upon pH. Corresponding quantum yields for formation of chlorine dioxide ranged from 0.43 t o 0.94, depending upon pH.
+
-
+
+
Introduction Chlorine dioxide is primarily used as an oxidizing agent in the treatment and bleaching of wood pulp, in water treatment, in textiles and wool treatment (bleaching),and in the treatment of flour and of food products. Chlorine dioxide can be produced by several methods. It can be prepared either by reduction of chlorates or by oxidation of chlorites. In large-scale commercial pulp bleaching operations, chlorine dioxide is produced by reducing chlorate in strong acid solution (Rapson, 1956,1958; Rapson and Wayman, 1949, 1954; Holst, 1945; Persson, 1945; Day and Fenn, 1949;Rosen, 1976; Engstrom and Norell, 1992; Engstrom and FalgBn, 1992). Variations of this process depend upon the type of acid (hydrochloric or sulfuric acid), operating acidity, and choice of reducing agent (methanol, sulfur dioxide, sodium chloride, hydrogen peroxide). All of these variables affect the chemistry of the process. Both final product and byproducts vary significantly depending on the set of operating conditions and choice of reducing agent. In water treatment, chlorine dioxide is produced by oxidation of chlorite. This operation is carried out on a much smaller scale and does not require equipment as extensive as that required for a chlorate reduction process. Usually a solution of sodium chlorite is mixed with a strong solution of chlorine in water yielding chlorine dioxide and chloride. Oxidizing agents such as ozone or persulfates can also be used instead of chlorine. Recently, several workers have studied an alternative way of producing chlorine dioxide. This process involves irradiation of an aqueous solution of sodium chlorite with ultraviolet light. The process is simple in terms of procedures and apparatus. The chlorine dioxide that is formed can be recovered in either gaseous or liquid phase, depending upon the purpose of the operation. This process requires only a single reactant, sodium chlorite, and the reaction can be terminated at any time by extinguishing the UV source. Several patents have been filed concerning methods of generating chlorine dioxide (Fisher, 1983,1984;Callerame,
* Author to whom correspondence should be addressed. 0888-5885/94/2633-1468$04.50/0
+
1989; Sekisui Plastics Co. Ltd., 1991) by photodecomposition of sodium chlorite with ultraviolet light. One of the potential drawbacks of this process is that chlorine dioxide decomposes when irradiated. Therefore chlorine dioxide must be recovered continuously from the reactor solution. Chlorine dioxide is produced in the aqueous phase, but can be gasified by sparging nitrogen or air through the solution. Variations of reactors have been studied. Ultraviolet emitting lamps can be placed inside (Callerame, 1989) or outside (Fisher, 1983,1984)the reactor. In the latter case, the reactor is constructed of quartz, Suprasil, or Vycor in order not to absorb ultraviolet light. Photodecomposition of Chlorine Dioxide. Little work has been published on the continuous photolysis of chlorine dioxide in aqueous solution. The following mechanism has been proposed for wavelengths between 300 and 436 nm (Bowen and Cheung, 1932):
hv
2C102(aq)
2C10 + 2 0
(1)
H,C10,
(2)
C10 + H,O C10 + H,C10,
+ HC1
(3)
+ C1- + 0, + 2H'
(4)
HClO,
and the overall reaction is BClO,(aq) + H,O
hu
C10;
Zika et al. (1984) proposed a different mechanism initiated by the same primary photoreaction, but leading to a different overall reaction and product distribution: 4ClO,(aq)
+ 2H,O
hv
2C10:
-+
+ C10- + C1- + 0, + 4H'
(5)
Recently Karpel Vel Leitner et al. (1992a)proposed an overall reaction for the decomposition of chlorine dioxide in water at 253.7 nm. 10C10, + 5H,O
hv
4C1- + 6C10;
0 1994 American Chemical Society
+ 3.50, + 10H'
(6)
Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994 1469 This global reaction is the result of a series of complex intermediate reactions initiated by ultraviolet radiation. The quantum yield of chlorine dioxide photodecomposition in water has been briefly investigated (Bowen and Cheung, 1932; Zika et al., 1984; Karpel Vel Leitner et al., 1992b). The first two of these research groups showed that the value of the quantum yield increases with decreasing wavelength. Recently, Zika et al. (1984) obtained a value of 1.4 at 300 nm and pH 7, while previous workers (Bowen and Cheung, 1932) reported a value of 1. These quantum yields are reported for the overall decomposition of chlorine dioxide, and it must be pointed out that primary quantum yield of decomposition is still unknown. No information on the quantum yield of chlorine dioxide decomposition at 253.7 nm was reported by any of these authors. Photodecomposition of Sodium Chlorite. Buxton and Subhani (1972a,b) have studied the photochemistry of oxychlorine ions. In particular, they focused their attention on the photodecomposition of aqueous solutions of chlorite ions (1972b). The products they observed in steady state photolysis experiments at pH 10 were C10-, Cl02, C1-, 02,and little C103-. In their experimental procedure, chlorine dioxide was removed from the reactor solution at frequent intervals. They proposed the following primary photodecomposition reactions at 254 nm:
hu
C10;
C10- + O('D)
hu
c10,-
(7)
c10 + 0-
(8)
(ClOL)*
(9)
hu
c10,-
Reaction 7 leads to the following secondary reactions:
O(lD) + H,O H,O,
+ C10-
-
-
H,O,
(10)
0, + C1- + H,O
(11)
A t 365 nm, some oxygen may also be formed in the (3P) state. This will lead to the following reactions:
-
+ cia, 0, + ci- + O ~ P ) o(3~ +)cio; c10,-
o(3~)
-
(12)
(13)
Reaction 8 leads to the following secondary reactions:
0- + H,O * OH + OHOH + C10;
-
+ OH-
C10,
2c10 * c1,0, Cl,O,
+ C10; + H,O
C10;
(14) (15) (16)
+ 2C10- + 2H'
(17)
In reaction 9, (ClOz-)* is a long-lived triplet state of chlorite ion. It reacts with another chlorite ion. (ClO;)*
+ C10;
C10,
+ C10- + 0-
(18)
Karpel Vel Leitner et al. (1992a,b) studied the photodecomposition of sodium chlorite in solution at 253.7 nm. They showed that chloride and chlorate were the major products of the decomposition. They proposed the following overall reaction:
hu
10C10;
6C1- + 4C10;
+ 40,
(19)
In their study, chlorine dioxide is considered as an intermediate product and is left in the reactor solution where it undergoes decomposition. In the second part of the study (1992b), they determined a value of 1 for the quantum yield of decomposition of sodium chlorite at natural pH. The quantum yield for the disappearance of ClOz- at 253.7 nm was found experimentally (Buxton and Subhani, 1972b) to be 0.50 at pH 10 and 0.79 at natural pH. The quantum yield for the production of chlorine dioxide was respectively 0.24 and 0.41 at these pH values. The present study was designed to achieve a better understanding of the production of chlorine dioxide by irradiation of sodium chlorite with ultraviolet light, especially in terms of stoichiometry, final stable products, and relative rates of production of products and byproducts. This understanding would enable us to determine the maximum yield of chlorine dioxide that can be achieved. We propose a mechanism which explains our experimental results and is consistent with previous work. Our experimental approach involved generating chlorine dioxide while simultaneously removing it from the reaction solution. This approach was aimed at minimizing further reaction of chlorine dioxide and thereby maximizing the ratio of chlorine dioxide to reacted chlorite. Experimental Section Photodecomposition of Chlorine Dioxide i n Solution. A. Photochemical Chamber Reactor. For this study we used a photochemical chamber reactor equipped with a maximum of 12 low-pressure mercury lamps. This homemade reactor is similar to a Rayonet Chamber Reactor RPR 100 (The Southern New England Ultraviolet Co., Branford, CT). The reactor was previously constructed and loaned to us by Dr. L. Tolbert of the Georgia Institute of Technology School of Chemistry. The reactor body is a cylindrical chamber 25.4 cm in diameter and 30 cm deep with 12 lamps aligned vertically at the wall. The chamber is cooled by a fan at the bottom, which allows for normal operating temperature between 25 and 30 "C. An external power supply contains the starters and a switch. The reactor can be operated with fewer lamps by removing some of them from their plugs. When fewer than 12 lamps were used, the remaining lamps were arranged symmetrically so as to ensure a uniform and symmetric exposure to the ultraviolet radiation. The walls are made of polished aluminum, which provides good reflection. When 12 lamps are used, the total power consumption (including the fan) approximately 240 W. We used germicidal low-pressure mercury lamps of type G8T5 (G.E. and Westinghouse), which provide an intense source of ultraviolet light. The mercury line radiates at 253.7 nm and accounts for approximately 90% of the energy radiated in the 250-600-nm region. We also used phosphor-coated low-pressure mercury lamps emitting at 300 nm, purchased from The Southern New England Ultraviolet Co. B. Reaction Vessel. The reaction vessel (shown in Figure 1, within the photochemical reaction chamber) is a 38-cm-long and 3.8-cm-diameter cylindrical vessel, designed specifically for this study. The vessel is made of quartz with a thickness of 1mm. This thickness ensures approximately 90 76 transmission down to a wavelength of 200 nm (Calvert, 1967; Koller, 1952). The vessel was suspended in the center of the photochemical chamber by a clamp.
1470 Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994 Spectrophotometer
m
a
J
5
-
3
Photochemical reactor
Figure 1. Experimental apparatus for the study of chlorine dioxide photodecomposition.
C. Chlorine Dioxide Solution. Aqueous solutions of chlorine dioxide were prepared by us using a simple procedure based on the reaction of sodium chlorate with oxalic acid in the presence of sulfuric acid (Masschelein, 1979). This reaction is reported to produce pure chlorine dioxide with no contamination by chlorine. The preparation is based on the following stoichiometry: H2S04+ 2NaC10,
+ H2C2O4 -, 2C10, + 2C0, + 2H20 + NazS04 (20)
The experimental apparatus consists of a 200-mL volumetric flask and three absorbers connected in series. The reactants were fed into the flask, and the mixture was stirred at ambient temperature. The gaseous products exited the reactor and traveled through the first absorber where chlorine was trapped in dibasic sodium phosphate and then through the second absorber where most of the chlorine dioxide was collected in distilled water at near freezing temperature. The third absorber contained distilled water and trapped chlorine dioxide which escaped the previous absorber. This setup allowed us to prepare a stock solution of 0.1 M chlorine dioxide (7 g/L). Stored in a brown bottle in a refrigerator, the chlorine dioxide solution is stable and can be used for up to 1week without detectable decomposition products. D. Irradiation Procedure. In the study of chlorine dioxide decomposition we irradiated a chlorine dioxide solution while continuously monitoring the absorbance and thus the concentration in chlorine dioxide (Figure 1). In a typical experiment, the reaction vessel was filled with 100 mL of chlorine dioxide solution; the solution was irradiated and pumped continuously through a 1-mm path length flow cell in a Milton Roy Spectronic 1201 spectrophotometer while the absorbance was monitored at 370 nm. The solution exiting the cell returned to the reactor. The movement of the solution also ensured its mixing. At 370 nm the extinction coefficient of chlorine dioxide in
solution is 104 M-l mm-I. The volume of solution circulating in the tubes at any given time was approximately 30 mL, and the pumping rate was 2 mL/s. All experiments were run between 23 and 25 "C. We used two wavelengths for irradiation, 300 and 253.7 nm. At 253.7 nm we examined the decomposition of chlorine dioxide as a function of the ultraviolet intensity by varying the number of lamps in the chamber. We used configurations with 12, 8, 6, 4, 2, and 0 lamps. With 0 lamps, the solution was exposed to ambient light only. Several experiments were run for each configuration. To improve accuracy, we used relatively high initial chlorine dioxide concentration (0.020-0.024 M), which ensured that the solution would capture a large fraction of the incident UV light. Photodecompositionof Sodium Chlorite in Solution at 253.7 nm. A. Choice of Ultraviolet Wavelength. Because UV light decomposes both chlorite and chlorine dioxide, an early task was to find a wavelength at which the coefficient of molecular extinction is minimum for chlorine dioxide (low absorption and therefore little tendency to react) and maximum for sodium chlorite (high absorption and therefore high tendency to react). Absorption spectra of aqueous solutions of chlorine dioxide and sodium chlorite were determined for the wavelength range of 230-370 nm using a Milton Roy Spectronic 1201spectrophotometer. The sodium chlorite spectrum was obtained from very stable absorbance of 0.0101MNaClOzatpH lland25"C. Thechlorinedioxide spectrum was obtained from 0.0017 M Cl02 at 25 OC. The extinction coefficient for chlorine dioxide shows a minimum at about 270 nm and that for chlorite a maximum at about 270 nm. Therefore 270 nm would be the optimal wavelength for these experiments. We chose to use low-pressure mercury lamps emitting at 253.7 nm since they are readily available. This wavelength is sufficiently near the optimal value of 270, and has been used in various patents and other references (Fisher, 1983,1984;Callerame, 1989;Buxton and Subhani, 1972b) claiming methods for generating chlorine dioxide by irradiation of sodium chlorite. B. Sodium Chlorite Solution. Sodium chlorite was prepared from Aldrich technical grade sodium chlorite (80%) following the procedure described by Peintler and Nagypal(1990). The carbonate content was precipitated with BaClz solution. The excess Ba2+ions were eliminated by adding a calculated amount of Na2S04 solution. The resulting NaClO2 was recrystallized twice from an 80% ethanol-water mixture at low temperature. This procedure yielded NaC102 with a chloride concentration of less than 0.7 7%. Sodium chlorite stock solution was prepared by diluting sodium chlorite crystals with distilled water, and was kept at room temperature in the dark to avoid any decomposition due to the ambient light. C. Experimental Apparatus and Procedure. The experimental apparatus consists of a quartz reaction vessel within a photochemical reaction chamber equipped with low-pressure mercury lamps emitting at 253.7 nm, and an absorption vessel. The quartz vessel and photochemical reaction chamber are those shown in Figure 1; however, the liquid recirculation tubing, pump and spectrophotometer were replaced by a gas absorption system. It consists of gas inlet and outlet tubing that extends through the top of the quartz vessel. Outside the vessel, the inlet tube connects to a nitrogen cylinder;inside it extends below the liquid surface where it connects to a glass frit positioned near the bottom of the vessel. The outlet tube extends from the gas space in the quartz vessel through the top and connects to a gas bubbler containing an absorbing liquid.
Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994 1471 The quartz vessel was filled with 70 mL of 0.1 M sodium chlorite solution and irradiated. Chlorine dioxide was sparged continuously from the reactor by a stream of nitrogen (UHP grade, Holox Ltd.) which then passed either directly to the vent or through the absorber which was filled with a 10% potassium iodide solution buffered at pH 9 with sodium phosphate. Chlorine dioxide reacted with iodide to give iodine, and nitrogen was vented to the hood. The procedure involved making several runs with different irradiation times, titrating for the various chemicalsin the reactor solution, and determining chlorine dioxide and chlorine production by titrating the absorber solution. The reactor solution was analyzed for hydroxide by titration with standard acid solution (Rand et al., 19751, for chloride by acidifying and titrating with standard silver nitrate solution (Rand et al., 1975), and for chlorate, chlorite, and hypochlorite by the titration procedure of Aieta et al. (1984). In this latter method, sodium thiosulfate is used to titrate iodine that is formed when potassium iodide is added to a sample that contains the chlorine-containing species. Sample pretreatments and pH adjustments are used to differentiate among the various chlorine-containing species. This method also detects chlorine dioxide; however, there was no chlorine dioxide left in the vessel because it was removed by the continuous, high nitrogen flow (4000 cm3/min). To further confirm there was no residual chlorine dioxide, we checked the final absorbance of the reactor solution at 370 nm and found it to be negligible. Iodine formed in the absorber was titrated with thiosulfate at pH 7 and pH 2, using the method of Aieta et al. (1984). The mols of chlorinedioxide and chlorine were then divided by the reactor solution volume in order to express their concentrations in terms of reactor volume. We examined the influence of several variables on the rate of production and on the stoichiometry of the reaction. These variables were nitrogen sparging flow rate, pH of the chlorite solution, and initial chlorite concentration in the reactor solution. Quantum Yield of Photodecomposition. In this study the quantum yield of photodecomposition is defined as the number of molecules decomposed (chlorine dioxide or sodium chlorite) divided by the number of photons absorbed by the solution. Therefore two separate measurements were required to calculate each quantum yield. The rate of photodecomposition of chlorine dioxide or sodium chlorite was obtained from our experiments. To determine the number of photons absorbed by the solution, we used chemical actinometry with potassium ferrioxalate. We followed the technique described by Hatchard and Parker (1956).
Results and Discussion Photodecomposition of Chlorine Dioxide in Aqueous Solution. Figure 2 show the results of the irradiation of chlorine dioxide at 253.7 and 300 nm. The absorbance of the reactor solution measured at 370 nm is recorded versus time of irradiation. In the experiment run without lamps the chlorine dioxide solution was exposed to ambient light. For safety reasons, the vessel was not sealed; therefore, a portion of the chlorine dioxide disappearance can be attributed to evaporation. This graph shows that ultraviolet radiation has a strong effect on the stability of chlorine dioxide. Irradiation causes a fast decomposition at a rate varying with the number of lamps used, and therefore with the flux of ultraviolet light absorbed by the solution.
2.0 l
1.5
umber of lamps in use
1 I
1.0 I I
0.5
0.0 0
20
40
60
80
100
Time (min) Figure 2. Photodecomposition of chlorine dioxide. Absorbance a t 370 nm versus time. Irradiation at 253.7 nm (a),300 nm (A),and ambient light (0). Flow cell with 1 mm path length.
Figure 2 shows that during about the first 10 min of each experiment the rate of decomposition was constant. We determined the rates based on the first 10 min of each experiment by fitting these data with straight lines. The slope of each line is the rate of decomposition in absorbance units per minute which we transformed to a rate in molar per minute using the extinction coefficient of chlorine dioxide at 370 nm (104 M-' mm-l). Table 1 summarizes the rate of decomposition corresponding to each configuration and each wavelength. QuantumY ield of Chlorine Dioxide Decomposition. Using rates of decomposition and ultraviolet flux determined by chemical actinometry, we calculated the quantum yield of chlorine dioxide decomposition. Rates of decomposition were corrected by subtracting the contribution due to the ambient light. The incident photon flux due to the ambient light was very small compared to the flux emitted by the lamps, and no correction was necessary. Table 2 shows that at 253.7 nm we found an average quantum yield of 0.42. At 300 nm we found a quantum yield of 1.38. This value is in very good agreement with the value of 1.40 given by Zika et al. (1984), and higher than the value of 1 given by Bowen and Cheung (1932). As stressed by Zika et al. (1984), this latter value may be underestimated due to the broad spectral region (316-270 nm) used. Photodecompositionof Sodium Chlorite in Aqueous Solution. A. Unbuffered Chlorite Solutions. A first set of experiments was conducted with unbuffered solutions of sodium chlorite. The initial chlorite concentration was between 0.09 and 0.1 M, with a chloride concentration below 0.001 M and a chlorate concentration below 0.001 M. After 60 min of irradiation the final pH was about 12.6,indicating the formation of OH-. The reactor solution was continuously sparged with nitrogen at 4000 cm3/min to remove chlorine dioxide. We titrated the reactor solution for C1-, ClO2-, OH-, ClOs-, and C10- and titrated the absorber solution for Cl02 and Cl2. Typical evolution of reactor species concentration is shown in Figure 3. Chlorite concentration is shown as moles of chlorite reacted per liter. Chlorate and hypochlorite are not shown but were titrated at the end of each run. No perchlorate was found in the reactor solution. Hydroxide was titrated and the pH of the reactor solution was calculated. Product concentrations (and consumed chlorite) are shown to be proportional to the irradiation time. After several runs we were able to determine an average rate of production for each product
1472 Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994 Table 1. Rate of Decomposition of Chlorine Dioxide ambient wavelength (nm) 0 no. of lamps rate of decompn (Mimin) X lo3 0.169
for Various Lamp Configurations 253.7 253.7 253.7 6 2 4 0.240 0.315 0.335
Table 2. Quantum Yield of Chlorine Dioxide Decomposition at 253.7 and 300 nm 253.7 wavelength (nm) ambient 253.7 4 no. of lamps 0 2 0.354 rate of photon abs (Mimin) X lo3 0 0.182 0.315 rate of ClOz decompn (M/min) X lo3 0.169 0.240 0.41 0.39 quantum yield of decompn 0.08 0.07
-
0.06
-
pH 12.6
-
I
5
253.7 12 0.551
253.7 8 0.583 0.427 0.44
253.7 6 0.394 0.335 0.42
0.06
____
253.7 8 0.427
300 8 1.009
~~~~~
/-
i
0.05
300 8 0.674 1.009 1.38
253.7 12 0.876 0.551 0.44
;
/’
E
/
.-
3
e
0.05
Y
E
0.04
2
e
0.03
h
0
Y
2
0.02
B
0.01
0.00 0
10
20
30
40
50
60
0
70
Time ( m i d Figure 3. Photodecomposition of sodium chlorite in solution. Evolution of reactor species and absorber species concentrations for unbuffered solutions. ClOz-reacted (e),ClOz (A),C1- (e),OH- W, Clz (0). Twelve lamps in use. Table 3. Irradiation of Sodium Chlorite. Average (Based on Three Runs) Rate of Production in the Case of Unbuffered Experiments. Twelve Lamps in Use rate of production product yield product (Mimin) X lo3 (M/M of chlorite reacted) ClOz- reacted 0.97 f 0.048 100 ClOz 0.58 f 0.030 60 f 3 c10.29 0.040 30 f 4 OH0.69 f 0.048 71 f 5 Clz 0.02 f 0.015 2f2 c100.05 f 0.005 5 f 0.5 4 f 0.5 c1030.04 f 0.005
*
and average proportion of each product compared to the amount of chlorite reacted. We calculated these proportions by forming the ratio of the product rate to the chlorite rate of dissociation. Data are summarized in Table 3. Hydroxide, chlorine dioxide, and chloride appear to be the major products of the decomposition of sodium chlorite. Chlorate and hypochlorite appear in lesser amounts. Chlorine appears in minor quantity, and the rate of production seems to fluctuate with time. B. Effect of pH on the Photodecomposition of Sodium Chlorite. We studied the photodecomposition of sodium chlorite solution buffered at various pH values. The solutions were prepared by first dissolving sodium chlorite in 0.2 M sodium phosphate solution yielding a solution with pH slightly above 9. Then we adjusted the pH to the desired value by slowly adding 1 M sulfuric acid. We studied solutions with pH ranging from 4 to 10. Below pH 4, sodium chlorite solutions tend to decompose rapidly even in the absence of light, by disproportionation. The pH value of 10 was achieved by buffering with 0.1 M sodium carbonate and adjusting the pH with sodium bicarbonate. In this set of experiments, the nitrogen flow rate was set at 4000 cm3/min and 1 2 lamps were operated. The initial chlorite concentration was between 0.09 and 0.1 M. After irradiation, we titrated for reactor species as before. The sum of chlorate and hypochlorite rates was determined
10
20
30
40
50
70
60
Time ( m i d Figure 4. Photodecomposition of sodium chlorite in solution. Evolution of reactor species and absorber species concentrations a t pH 10. ClOz- reacted ( O ) ,ClOz (A),CI- (e),Clz (0). Twelve lamps in use. Table 4. Irradiation of Sodium Chlorite. Average (Based on Three runs) Rate of Production and Product Yields in the Case of Buffered Experiments. Twelve Lamps in Use species
pH4 pH5 pH6 pH7 Rate of Production C102-reacted 1.56 1.79 1.92 1.48 1.00 1.15 1.18 0.84 c102 Cl0.42 0.47 0.53 0.47 CI 0.05 0.07 0.02 0.03 C103-andCIO- 0.05 0.04 0.19 0.10 C102-reacted ClOZ
100
64 c127 c12 3 C103-andC10- 3
p H 8 pH9 pH 10 average (M/min) X lo3 1.15 0.95 0.90 0.67 0.59 0.54 0.38 0.24 0.21 0.02 0.00 0.01 0.046 0.11 0.14
Product Yield (M/M of Chlorite Reacted) 100 100 100 100 100 100 64 61 57 59 62 60 26 27 32 33 26 23 4 1 2 2 0 1 2 10 7 4 12 15
100 61 28 2
8
by difference to achieve a chlorine balance. The reactor solution showed a very slight change in pH. Figure 4 shows the evolution of reactor products during a run at pH 10. Table 4 shows average rates of production and product yields at various pH conditions. Chlorine dioxide and chloride yields are reasonably constant, and pH does not seem to have a significant effect on them. However the rate at which sodium chlorite reacts is affected by the pH of the reactor solution. A maximum rate appears at a pH value of 6. Figure 5 shows the rate of formation of chlorine dioxide versus pH of the chlorite solution. Compared to this curve, the rate of formation in the case of unbuffered chlorite solutions seems to be consistent, as the pH was above 1 2 throughout most of the unbuffered experiments. C. Stoichiometry and Mechanism of the Photodecomposition of Sodium Chlorite. Our results show that when we varied pH, the product distribution, hence the stoichiometry of the overall reaction to the major products, C1-, C102, and OH-, remained reasonably constant even though the overall reaction rate varied significantly. This result provides indirect evidence about the mechanistic steps of the process: that the stoichiometry of the overall reaction is fixed by the elementary steps of the mechanism,
Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994 1473 c
-3. -2 g
0.0012 0.0011
-
OH + C10;
m
:
C10,
+ OH-
(15)
C10- is also decomposed by ultraviolet light as shown by Buxton and Subhani (1972a). They proposed
0.001 1
- + - + hv
0.0009 i
C10-
. YI
C1-
O('D)
(22)
ci-
o(3~)
(23)
hu
cio-
0.0007
E
'
0.0006
..
0.0004
ZzIIzl
hv
Unbuffered experiment (average)
Y
0.0005
ON
3
4
5
6
7
8
9
1
0
1
1
pH of reactor solution Figure 5. Rate of formationof chlorine dioxide at various pH values and in the case of unbuffered solutions.
and that the steps are interdependent in such a way that they do not allow for independent production of the products, C1-, C102, and OH-. Karpel Vel Leitner et al. (1992a,b) studied the photodecomposition of chlorine dioxide and sodium chlorite in aqueous solution by irradiation at 253.7 nm. They used a similar reactor and low-pressuremercury lamps; however in their work chlorine dioxide was not sparged from the reactor, but was subject to decomposition. In a first set of experiments, they studied the decomposition of chlorine dioxide and determined that chloride and chlorate were the major decomposition products. They reported yields of 0.4 mol of chloride and 0.6 mol of chlorate per mole of reacted chlorine dioxide. In a second set of experiments they studied the decomposition of sodium chlorite. Chloride and chlorate were also the major decomposition products, reported at yields of 0.6 mol of chloride and 0.4 mol of chlorate per mole of reacted chlorite. We will use our results and the results of others in proposing a stoichiometry and a mechanism for the photodecomposition of chlorite. We will show that our results are consistent with those of Karpel Vel Leitner et al. (1992a) if we assume that chlorate forms only from chlorine dioxide. We will refer to other studies (Karpel Vel Leitner et al., 1991, 1992a,b; Buxton and Subhani, 1972a,b) to explain the presence of hypochlorite. We propose the following overall stoichiometry for complete conversion of chlorite: 3ClO;
+ H,O
hv
C1- + 2C10, + 20H- + 0.50, (21)
To explain the stoichiometry, we examine reaction steps which have been previously proposed. According to Buxton and Subhani (1972b), the primary photodecomposition of chlorite ion may occur in several ways:
hv
c10,(ClO;)*
+ C1OL
-
hv
C10;
C10,
+ C10- + 0-
C10- + O('D)
hv
c10,-
(ClO,-)*
c10 + 0-
(9)
(18) (7) (8)
followed by secondary reactions, 0- + H,O
OH- + OH
(14)
c10-
c1+ 0-
(24)
Our proposed stoichiometry can be explained by the addition of reactions 9,18,14,15, and 22 (or 23), the sum of which equals reaction 21. On the basis of our results we believe that reaction 9 is more likely to occur than reaction 8 since C10 would further react, yielding chlorate. Reaction 7 has been shown by Buxton and Subhani to be unimportant at 253.7 nm. Similarly we suspect that reactions 22 and 23 are faster than 24 since C1 would further react (in a complicated pattern involvingchlorite), producing chlorate. It has also been shown by Karpel Vel Leitner et al. in another article (1991)that the reaction between hypochlorite and chlorite in basic solution produces chlorate. However, our results suggest that this reaction, like reactions 7,8, and 24, is of lesser importance than the reactions we included in the mechanism. Our overall stoichiometry is supported by the overall photodecomposition reactions reported by Karpel Vel Leitner et al. (1992a). For the decomposition of chlorine dioxide they proposed the following reaction: 10C10,
+ 5H,O
hv
+ 6ClOi + 3.50, + 10H'
4C1-
(6)
For the decomposition of chlorite, they proposed
hv
10C10;
6C1-
+ 4C10; + 40,
(19)
If we treat these two reactions as simultaneous algebraic equations, we eliminate chlorate by multiplying equation 19 by 3 and equation 6 by 2, and subtracting one equation from the other. The result of this calculation is 3C10;
+ 2H'
-
C1- + 0.50,
+ 2C10, + H,O
(25)
If we combine this equation with H,O
-
H+ + OH-
(26)
and eliminate H+, we obtain equation 21. Assuming the proposed mechanism (reactions 9,18,14, 15,22,or 23) for formation of chlorine dioxide and assuming chlorate forms from chlorine dioxide by reaction 6, we can estimate the distribution of chlorine dioxide, chloride, and hydroxide we might have observed in our experiments if chlorine dioxide had been completely sparged from the reactor before it reacted. Table 5 (column 2) shows the experimental distribution. To eliminate chlorate from the products in Table 5, we use the stoichiometry of reaction 6. This reaction shows that the 4 mol of chlorate would be produced from 410.6 = 6.7 mol of chlorine dioxide. Therefore, if chlorate had not formed, the product distribution would contain an additional 6.7 mol of chlorine dioxide. Reaction 6 also predicts that the products would contain an additional 0.4 X 6.7 = 2.7 mol of chloride and would contain 6.7 mol less of hydroxide ions.
1474 Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994
D. Quantum Yield of Sodium Chlorite Photodecomposition and Chlorine Dioxide Production. To determine the quantum yield of sodium chlorite decomposition, we divided the rate of decomposition of sodium chlorite by the absorbed photon flux determined earlier. Sodium chlorite decomposition experiments were run with a reactor configuration utilizing 12 lamps. The absorbed photon flux is in that case 1,4606 X lo4 mol of photon/s. Table 6 shows the quantum yield of photodecomposition of chlorite and the quantum yield of production of chlorine dioxide for various pH values. These results are in very good agreement with the study of Karpel Vel Leitner et al. (199213). They found a quantum yield of photodecomposition of sodium chlorite equal to 1for a pH between 7 and 8. We found a value of 1.18 at pH 7 and 0.92 at pH 8. Our results are also in good agreement with the results given by Buxton and Subhani (1972b). A t 253.7 nm they found a quantum yield of chlorite decomposition equal to 0.54 at pH 10 and 0.9 at natural pH. For the quantum yield of chlorine dioxide production, they determined a value of 0.24 at pH 10 and 0.41 at natural pH. E. Effect of Nitrogen Flow Rate on the Reaction. The effect of the nitrogen flow rate used in sparging product gases from solution, has been investigated using the reactor equipped with 1 2 lamps and unbuffered chlorite solution with a concentration of 0.095 M. Flow rates ranged between 0 and 5000 cm3/min. Table 7 shows the rates of production and product yields for several values of nitrogen flow rate. The table shows that, at zero flow rate, no chlorine dioxide is observed after a sufficient time of irradiation. As flow rate increases, chlorine dioxide and hydroxide increase while chloride and chlorate plus hypochlorite decrease substantially. These results are consistent with previous discussion and show that at low flow rates our results suggest the stoichiometry observed by Karpel Vel Leitner et al. (1992a). A t high flow rate (as chlorine dioxide is more rapidly removed) the stoichiometry approaches that of reaction 21. The sparging rate plays an important part in the production of chlorine dioxide. If it is too low, chlorine dioxide is decomposed due to a large residence time in the reactor. The decomposition of chlorine dioxide absorbs part of the ultraviolet light which is therefore not available for the decomposition of chlorite. This would explain the decrease in the rate of decomposition of chlorite at low nitrogen flow rate. If the flow rate is too high, some of the reactor solution might be removed by entrainment. F. Effect of Initial Sodium Chlorite Concentration. The effect of initial chlorite concentration was investigated
Table 5. Experimental Product Yields, Corrected Experimental Product Yields, and Theoretical Product Yields Based on Reaction 21, in the Case of Unbuffered Reactor Solutions ~
expert1 product yields unbuffered coir for theor product yields sPecies soln C102 decompn based on reaction 21 ClOz- reacted 100 100 100 ClOZ 60 f 3 67 66.7 c130 f 4 32 33.3
*
OH-
71 5 2*2 5 f 0.5 4 f 0.5
Clz c10c103-
64
66.7
Table 6. Quantum Yield of Photodecomposition of Sodium Chlorite and Quantum Yield of Production of Chlorine Dioxide at 254 nm, for Various pH Conditions PH
product
average rate (M/min) X lo3
quantum yield
unbuffered
ClOz produced Cl02- reacted C102 produced ClOz- reacted ClOz produced C102- reacted C102 produced ClOz- reacted ClOz produced ClOz- reacted Cl02 produced ClOz- reacted C 1 0 ~produced C102- reacted ClOz produced ClOz- reacted
0.583 0.967 1.000 1.556 1.149 1.789 1.176 1.922 0.839 1.482 0.674 1.148 0.586 0.948 0.543 0.900
0.47 0.77 0.80 1.24 0.92 1.43 0.94 1.53 0.67 1.18 0.54 0.92 0.47 0.76 0.43 0.72
4 5 6 7 8 9 10
To eliminate the 5 mol of hypochlorite from the products in Table 5, we assume that hypochlorite would eventually react by reaction 22 or 23, producing 5 mol of chloride. Therefore, this 5 mol would be added to the product distribution. We did not attempt to account for the chlorine which we found in the absorber samples, because of the large uncertainty associated with its concentration and the large percentage deviation we observed in repeated experiments. The sums of all of the adjustments to the product distribution are -4 mol of chlorate, -5 mol of hypochlorite, +6.7 mol of chlorine dioxide, +7.7 mol of chloride, and -6.7 mol of hydroxide. Table 5 gives a comparison between experimental coefficients, corrected experimental coefficients, and theoretical stoichiometric coefficients based on reaction 21.
Table 7. Irradiation of Sodium Chlorite. Effect of Nitrogen Flow Rate on Rates of Production and Product Yields. Twelve Lamps in Use. Unbuffered Reactor Solutions nitrogen flow rate (cm3/min) species
0
500
1000
2000
3000
Rate of Production (Mimin) x ClOz- reacted ClOZ c1-
0.61 0 0.26
OHCl2 C103- and C10-
NMa NM 0.29 and 0.06
C102- reacted ClOZ c1-
100 0 43
OH-
NM NM 48 and 10
0.70 0.14 0.29 0.29 0.06 0.15
0.77 0.26 0.32 0.43 0.06 0.17
0.81 0.41 0.25 0.53 0.04 0.06
4000
5000
lo3 0.87 0.48 0.27 0.60 0.03 0.06
0.97 0.58 0.29 0.69 0.02 0.04 and 0.05
0.95 0.57 0.28 0.70 0.02 0.06
Product Yield (M/M of Chlorite Reacted)
Clz C l o g and C100
NM = not measured.
100 20 42 41 8 22
100 33 42 56 8 9
100 51
100 55
32 66 5 7
32 69 3 7
100 60 29 71 2 4 and 5
100 60
30 74 2 6
Ind. Eng. Chem. Res., Vol. 33, No. 6, 1994 1475 using unbuffered sodium chlorite solutions and the photochemical reactor equipped with 12 lamps. The nitrogen flow rate was held constant at 4000 cm3/min. The initial concentrations used were 0.01, 0.04, and 0.09 M. Product yields were not affected by the change in the initial concentration. At the conditions of this study, the initial concentration does seem to have a significant effect on rates of production and rate of chlorite photodecomposition.
Conclusions In studies of photodecomposition of chlorine dioxide by irradiation with ultraviolet light, we determined the quantum yield of photodecomposition of chlorine dioxide at 253.7 and 300 nm. At 300 nm our results agreed with the value 1.4 given in the literature (Zika et al., 1984). At 253.7 nm we determined a quantum yield of photodecomposition equal to 0.44 at 25 OC. In studies of the reaction of photodecomposition of sodium chlorite in aqueous solution with rapid removal of chlorine dioxide by sparging with nitrogen, we showed chloride and chlorine dioxide to be the major products of the photodecomposition. We propose the followingoverall reaction: 3C10;
+ H,O
hu
C1- + 2C10,
+ 20H- + 0.50,
On the basis of our work and the work of several authors, we conclude that chlorate is not formed directly as a product of the photodecomposition of chlorite, but by decomposition of chlorine dioxide produced from chlorite. Although some hypochlorite may be formed as an intermediate, it is also further decomposed to chloride and oxygen. At natural pH, we found a quantum yield of chlorite decomposition equal to 0.77 and a quantum yield of chlorine dioxide production equal to 0.47. These values are in good agreement with those found in the literature (Buxton and Subhani, 1972b). The rate of photodecomposition of chlorite increases with decreasing pH. Product yields remained constant for all values of pH.
Acknowledgment We gratefully acknowledgeEka Nobel, Inc., for financial support of this work. We thank Dr. M. Fazlul Hoq for technical assistance, especially in establishing analytical procedures and chemical purification methods, and Dr. Laren Tolbert for use of the photochemical chamber and for helpful discussions and advice.
Nomenclature (ClOz-)* = chlorite ion in excited state O(lD) = oxygen atom in excited state O(3P) = oxygen atom in ground state hv = photon
Literature Cited Aieta,E. M.;Roberts, P. V.; Hernandez, M. Determination of Chlorine, Chlorite, and Chlorate in Water. J. Am. Water Works Assoc. 1984, 76 (l), 64.
Bowen, E. J.; Cheung, W. M. The Photodecomposition of Chlorine Dioxide Solution. Proc. Chem. SOC.1932, 1200. Buxton, G. V.; Subhani, M. S. Radiation Chemistry and Photochemistry of Oxychlorine Ions. Part 2.-Photodecomposition of Aqueous Solutions of Hypochlorite Ions. J. Chem. Soc., Faraday Trans. 1 1072a, 68,958. Buxton, G. V.; Subhani, M. S. Radiation Chemistry and Photochemistry of Oxychlorine Ions. Part 3.-Photodecomposition of Aqueous Solutions of Chlorite Ions. J.Chem. Soc., Faraday Trans. 1 1972b,68,970. Callerame, J. Process for the Production of Chlorine Dioxide. US Pat. 4,874,489, 1989. Calvert, J. G.; Pitts Jr., J. N. Photochemistry; John Wiley & Sons Inc.: New York, NY, 1967; p 748. Day, G. A.; Fenn, E. F. Process for Producing Chlorine Dioxide. US Pat. 2,484,402, 1949. Engstrom, J.; Norell, M. Process for the Production of Chlorine Dioxide. US Patent No. 5091166, Feb 25, 1992. Engstrom, J.; FalgBn, H. Process for the Production of Chlorine Dioxide. US Patent No. 5091167, Feb 25, 1992. Fisher, R. P. Method for Generating Chlorine Dioxide Gas. US Pat, 4,414,180, 1983. Fisher, R. P. Photochemical Method for Generating Chlorine Dioxide Gas. US Pat. 4,456,411, 1984. Hatchard, C. G.; Parker, C. A. ANew Sensitive Chemical Actinometer. 11. Potassium Ferrioxalate as a Standard Chemical Actinometer. Proc. R. SOC.l956,253A, 518. Holst, T. G. Processes for Producing Chlorine Dioxide. US Pat. 2,373,830, 1945. Karpel Vel Leitner, N.; De Laat, J.; Dor6, M.; Suty, H. Kinetics of the Reaction Between Chlorine and Chlorite in Dilute Aqueous Solution. Environ. Technol. 1991, 12, 477. Karpel Vel Leitner, N.; De Laat, J.; Dor6, M. Photodecomposition of Chlorine Dioxide and Chlorite by U.V.-irradiation-Part I. Photoproducts. Water Res. 1992a, 26, 1655. Karpel Vel Leitner, N.; De Laat, J.; Dor6, M. Photodecomposition of Chlorine Dioxide and Chlorite by U.V.-irradiation-Part 11. Kinetic Study. Water Res. 1992b, 26, 1665. Koller, L. R. Ultraviolet Radiation; John Wiley & Sons Inc.: New York, NY, 1952; pp 140-150. Masschelein, W. J. Chlorine Dioxide; Ann Arbor Science Publications: Ann Arbor, MI, 1979. Peintler, G.; Nagypal, I. Kinetics and Mechanism of the Reaction Between Chlorite Ion and Hypochlorous Acid. J. Phys. Chem. 1990, 94, 2954. Persson, S. H. Chlorine Dioxide Manufacture. US Pat. 2,376,935, 1945. Rand, M. C.; Greenberg, A. E.; Taras, M. J. Eds. Standards Methods for the Examination of Water and Wastewater,14thed.;American Public Health Association: Washington, DC, 1975; pp 273 and 306. Rapson, W. H. The Mechanism of Formation of Chlorine Dioxide from Sodium Chlorate. TAPPZ 1956,39, 554. Rapson, W. H. Recent Developments in the Manufactureof Chlorine Dioxide. Can. J. Chem. Eng. 1958, 36, 262. Rapson, W. H.; Wayman, M. Method for Removing Chlorine from Mixtures Containing Chlorine Dioxide and Chlorine. US Pat. 2,481,241, 1949. Rapson, W. H.; Wayman, M. A Continuous Process for the Manufacture of Chlorine Dioxide. Pulp Pap. Mag. Can. 1954,55 (61, 92. Rosen, H. J. Method of Simultaneously Producing Chlorine Dioxide and a Sulfate Salt. US Pat. 3,933,988, 1976. Sekisui Plastics Co. Ltd. Preparing Chlorine Dioxide by Irradiating UV on Chlorite Aqueous Solution. Jap. Pat. 091375, 1991. Zika, R. G.; Moore, C. A.; Gidel, L. T.; Cooper, W. J. Sunlight-induced Photodecomposition of Chlorine Dioxide. Water Chlorination: Chem. Environ. Impact Health Eff.,Proc. Conf. 1984,5th, 1041.
Received for reuiew July 26, 1993 Revised manuscript received February 22, 1994 Accepted March 7, 1994' Abstract published in Advance ACS Abstracts, April 15, 1994.
